Bucket Depth Peeling. Liu, F., Huang, M., Liu, X., & Wu, E. In SIGGRAPH 2009: Talks, of SIGGRAPH '09, pages 50:1--50:1, New York, NY, USA, 2009. ACM. doi abstract bibtex Efficient rendering of multi-fragment effects has long been a great challenge in computer graphics. The classical depth peeling algorithm [Everitt 2001] provides a simple but robust solution by peeling off one layer per pass, but multi rasterizations will lead to performance bottleneck for large and complex scenes. The k-buffer [Bavoil et al. 2007] captures k fragments in a single pass but suffers from read-modify-write(RMW) hazards which can be alleviated by multi passes [Liu et al. 2006]. Our approach exploits multiple render targets (MRT) as bucket array per pixel. Fragments are scattered into different buckets and sorted by a bucket sort. We describe two efficient schemes to reduce collisions when multiple fragments are routed to the same bucket. Our algorithm shows up to 32 times speedup to depth peeling especially for large scenes and the results are visually faithful. Also it has no requirement of pre-sorting geometries or post-sorting fragments, and is free of RMW hazards.
@inproceedings{Liu:ea:2009,
abstract = {Efficient rendering of multi-fragment effects has long been a great challenge in computer graphics. The classical depth peeling algorithm [Everitt 2001] provides a simple but robust solution by peeling off one layer per pass, but multi rasterizations will lead to performance bottleneck for large and complex scenes. The k-buffer [Bavoil et al. 2007] captures k fragments in a single pass but suffers from read-modify-write(RMW) hazards which can be alleviated by multi passes [Liu et al. 2006]. Our approach exploits multiple render targets (MRT) as bucket array per pixel. Fragments are scattered into different buckets and sorted by a bucket sort. We describe two efficient schemes to reduce collisions when multiple fragments are routed to the same bucket. Our algorithm shows up to 32 times speedup to depth peeling especially for large scenes and the results are visually faithful. Also it has no requirement of pre-sorting geometries or post-sorting fragments, and is free of RMW hazards.},
acmid = {1598040},
added-at = {2016-02-04T20:35:14.000+0100},
address = {New York, NY, USA},
articleno = {50},
author = {Liu, Fang and Huang, Meng-Cheng and Liu, Xue-Hui and Wu, En-Hua},
biburl = {http://www.bibsonomy.org/bibtex/217d6cc5bec9723481a4157726caf7c1e/ledood},
booktitle = {SIGGRAPH 2009: Talks},
description = {Bucket depth peeling},
doi = {10.1145/1597990.1598040},
interhash = {f78b8d20c2fd4b298ee7bf7d857beef7},
intrahash = {17d6cc5bec9723481a4157726caf7c1e},
isbn = {978-1-60558-834-6},
keywords = {depth-peeling},
location = {New Orleans, Louisiana},
numpages = {1},
pages = {50:1--50:1},
publisher = {ACM},
series = {SIGGRAPH '09},
timestamp = {2016-04-29T15:27:14.000+0200},
title = {Bucket Depth Peeling},
year = 2009
}
Downloads: 0
{"_id":"p7WboLjNhQR57jWRX","bibbaseid":"liu-huang-liu-wu-bucketdepthpeeling-2009","downloads":0,"creationDate":"2016-06-28T06:36:14.518Z","title":"Bucket Depth Peeling","author_short":["Liu, F.","Huang, M.","Liu, X.","Wu, E."],"year":2009,"bibtype":"inproceedings","biburl":"http://www.bibsonomy.org/bib/author/Fang Liu?items=1000","bibdata":{"bibtype":"inproceedings","type":"inproceedings","abstract":"Efficient rendering of multi-fragment effects has long been a great challenge in computer graphics. The classical depth peeling algorithm [Everitt 2001] provides a simple but robust solution by peeling off one layer per pass, but multi rasterizations will lead to performance bottleneck for large and complex scenes. The k-buffer [Bavoil et al. 2007] captures k fragments in a single pass but suffers from read-modify-write(RMW) hazards which can be alleviated by multi passes [Liu et al. 2006]. Our approach exploits multiple render targets (MRT) as bucket array per pixel. Fragments are scattered into different buckets and sorted by a bucket sort. We describe two efficient schemes to reduce collisions when multiple fragments are routed to the same bucket. Our algorithm shows up to 32 times speedup to depth peeling especially for large scenes and the results are visually faithful. Also it has no requirement of pre-sorting geometries or post-sorting fragments, and is free of RMW hazards.","acmid":"1598040","added-at":"2016-02-04T20:35:14.000+0100","address":"New York, NY, USA","articleno":"50","author":[{"propositions":[],"lastnames":["Liu"],"firstnames":["Fang"],"suffixes":[]},{"propositions":[],"lastnames":["Huang"],"firstnames":["Meng-Cheng"],"suffixes":[]},{"propositions":[],"lastnames":["Liu"],"firstnames":["Xue-Hui"],"suffixes":[]},{"propositions":[],"lastnames":["Wu"],"firstnames":["En-Hua"],"suffixes":[]}],"biburl":"http://www.bibsonomy.org/bibtex/217d6cc5bec9723481a4157726caf7c1e/ledood","booktitle":"SIGGRAPH 2009: Talks","description":"Bucket depth peeling","doi":"10.1145/1597990.1598040","interhash":"f78b8d20c2fd4b298ee7bf7d857beef7","intrahash":"17d6cc5bec9723481a4157726caf7c1e","isbn":"978-1-60558-834-6","keywords":"depth-peeling","location":"New Orleans, Louisiana","numpages":"1","pages":"50:1--50:1","publisher":"ACM","series":"SIGGRAPH '09","timestamp":"2016-04-29T15:27:14.000+0200","title":"Bucket Depth Peeling","year":"2009","bibtex":"@inproceedings{Liu:ea:2009,\n abstract = {Efficient rendering of multi-fragment effects has long been a great challenge in computer graphics. The classical depth peeling algorithm [Everitt 2001] provides a simple but robust solution by peeling off one layer per pass, but multi rasterizations will lead to performance bottleneck for large and complex scenes. The k-buffer [Bavoil et al. 2007] captures k fragments in a single pass but suffers from read-modify-write(RMW) hazards which can be alleviated by multi passes [Liu et al. 2006]. Our approach exploits multiple render targets (MRT) as bucket array per pixel. Fragments are scattered into different buckets and sorted by a bucket sort. We describe two efficient schemes to reduce collisions when multiple fragments are routed to the same bucket. Our algorithm shows up to 32 times speedup to depth peeling especially for large scenes and the results are visually faithful. Also it has no requirement of pre-sorting geometries or post-sorting fragments, and is free of RMW hazards.},\n acmid = {1598040},\n added-at = {2016-02-04T20:35:14.000+0100},\n address = {New York, NY, USA},\n articleno = {50},\n author = {Liu, Fang and Huang, Meng-Cheng and Liu, Xue-Hui and Wu, En-Hua},\n biburl = {http://www.bibsonomy.org/bibtex/217d6cc5bec9723481a4157726caf7c1e/ledood},\n booktitle = {SIGGRAPH 2009: Talks},\n description = {Bucket depth peeling},\n doi = {10.1145/1597990.1598040},\n interhash = {f78b8d20c2fd4b298ee7bf7d857beef7},\n intrahash = {17d6cc5bec9723481a4157726caf7c1e},\n isbn = {978-1-60558-834-6},\n keywords = {depth-peeling},\n location = {New Orleans, Louisiana},\n numpages = {1},\n pages = {50:1--50:1},\n publisher = {ACM},\n series = {SIGGRAPH '09},\n timestamp = {2016-04-29T15:27:14.000+0200},\n title = {Bucket Depth Peeling},\n year = 2009\n}\n\n","author_short":["Liu, F.","Huang, M.","Liu, X.","Wu, E."],"key":"Liu:ea:2009","id":"Liu:ea:2009","bibbaseid":"liu-huang-liu-wu-bucketdepthpeeling-2009","role":"author","urls":{},"keyword":["depth-peeling"],"downloads":0,"html":""},"search_terms":["bucket","depth","peeling","liu","huang","liu","wu"],"keywords":["depth-peeling"],"authorIDs":[],"dataSources":["fpXCfrFZJBDNs2m89"]}